Let's assume that we have the following scenario happening in a database server (also assume that this is the only thing happening in it):
5 table scans started at the same time, hitting different databases but the table structure and number of rows are all identical (roughly 200 million rows).
When I do a mysqladmin extend -r -i 10 | grep Handler_read_rnd_next
It shows that this variable is being updated. The question is since there are 5 threads running, are all of them hitting the Handler* subroutine one after the other? If so, would this slow down the whole table scanning process?
I tried to test this by running 1 table scan and it took about 1 hour to finish. And if I try to run 2 table scans together, it takes about 4 1/2 hours to finish.